get_relevant_context
Searches the team's context library and returns the most relevant chunks for a given prompt. Use this to gather supporting context before generating or refining content.
Parameters
| Name | Type | Required | Default | Description |
|---|---|---|---|---|
prompt | string | Yes | — | Search string |
project_id | string/uuid | No | — | Project ID to scope search |
collection_ids | array | No | — | Collection IDs to scope search |
context_rag_ids | array | No | — | Chunk IDs to exclude (for pagination) |
Input Schema
{
"type": "object",
"required": [
"prompt"
],
"properties": {
"prompt": {
"type": "string",
"description": "Search string"
},
"project_id": {
"description": "Project ID to scope context search"
},
"collection_ids": {
"description": "Collection IDs to scope context search"
},
"context_rag_ids": {
"description": "Previously returned chunk IDs to exclude (pagination)"
}
}
}Output Schema
{
"type": "object",
"properties": {
"context_rag_ids": {
"type": "array",
"items": {
"type": "string",
"format": "uuid"
},
"description": "Chunk IDs returned. Pass back to exclude from future searches."
},
"context_item_ids": {
"type": "array",
"items": {
"type": "string",
"format": "uuid"
},
"description": "Parent context item IDs that the chunks belong to."
},
"relevant_context": {
"type": "string",
"description": "Concatenated relevant context text from matched chunks"
}
}
}Instructions
Searches the team's context library and returns the most relevant chunks for a given prompt. Use this to gather supporting context before generating or refining content.
Pass a descriptive prompt to get the best matching context. Optionally scope results to specific collections or a project. Use context_rag_ids to exclude previously retrieved chunks when paginating for additional context.
Returns up to 10 context chunks per call.